A Small‐Sample Comparison of Estimators in the EU‐MGF Approach to Decision Making
比较了非参数经验矩生成函数与参数估计量在决策中的小样本表现，考察了样本量、风险厌恶程度和模型近似程度的影响，给出了不同情形下的选择建议。
Abstract Estimation of the moment‐generating function lies at the core of the exponential utility—moment‐generating function approach to decision making. The small sample performances of the nonparametric empirical moment‐generating function and a parametric competitor have been examined under a variety of situations defined by the sample size, the level of risk aversion, and the degree to which the assumed parametric model approximates reality. Conditions under which each estimator would be preferred are obtained. Neither approach can be recommended unequivocally in all situations.